Targeting the Gpr125-Gper1 axis: a novel osteogenic signaling pathway with therapeutic potential against osteoporosis

Abstract

Background

Osteoporosis, characterized by reduced bone mass and increased fracture risk, underscores the urgent need for novel anabolic therapies. G protein-coupled receptors (GPCRs) are major drug targets, but the functions of adhesion GPCRs (aGPCRs) in bone remain largely unexplored. G-protein coupled receptor 125 (Gpr125) is an orphan aGPCR, and its role in osteoblast-mediated bone formation is entirely unknown.

Methods

We employed lentiviral-mediated knockdown and overexpression of Gpr125 in primary mouse calvarial osteoblasts and bone marrow stromal cells (BMSCs). Osteogenic and adipogenic differentiation were assessed by staining and marker analysis. Transcriptomic profiling (RNA-seq) and pathway analysis were used to identify downstream mechanisms, validated by rescue experiments with Gper1 overexpression, PI3K/AKT inhibitors (LY294002), and Wnt/β-catenin activation (Wnt3a CM). The osteogenic role of Gpr125 was tested in ovariectomized and aged mouse osteoporosis models via osteoblast-targeted adeno-associated virus (AAV) delivery.

Results

Gpr125 was highly expressed in osteoblasts, peaking during differentiation. Its knockdown severely impaired osteogenesis while promoting adipogenesis in vitro. Conversely, its overexpression enhanced bone formation. RNA-seq identified G protein-coupled estrogen receptor 1 (Gper1) as the key downstream target. We defined a novel signaling axis where Gpr125 upregulates Gper1, which activates PI3K/AKT signaling, leading to β-catenin stabilization and osteogenic transcription. Rescue experiments established a strict hierarchy: Gper1 overexpression fully rescued the osteogenic defect caused by Gpr125 loss, but not Gpr125 expression itself. PI3K inhibition blocked Gpr125-induced β-catenin activation and osteogenesis. β-catenin activation partially rescued osteogenesis but failed to restore upstream signaling. Critically, osteoblast-specific Gpr125 overexpression in vivo effectively ameliorated bone loss and reduced marrow adiposity in both postmenopausal and senile osteoporosis mouse models.

Conclusions

Our study unveils a complete Gpr125-Gper1-PI3K/AKT-β-catenin signaling axis essential for osteoblast differentiation and bone formation. This work identifies the aGPCR Gpr125 as a novel positive regulator of bone anabolism and proposes the Gpr125-Gper1 axis as a promising therapeutic target for developing new treatments against osteoporosis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-08337-1.

Introduction

Osteoporosis is a prevalent metabolic bone disorder that manifests clinically through progressive bone mass reduction and microarchitectural deterioration, culminating in compromised bone strength and heightened fracture susceptibility [1–5]. This condition exhibits an insidious progression, frequently remaining clinically silent until the occurrence of fragility fractures - a characteristic that has earned osteoporosis its designation as a “silent disease” [4]. The asymptomatic nature of bone loss results in most patients remaining undiagnosed until the first fracture. Post-fracture complications severely impair mobility and quality of life, often triggering a cascade of secondary fractures that substantially elevate mortality rates [6], thereby imposing profound burdens on both global healthcare systems and socioeconomic structures. Consequently, elucidating the molecular pathogenesis of osteoporosis and developing targeted therapeutic strategies are critical for developing safe and effective treatment modalities.

Current pharmacotherapeutic approaches predominantly target bone resorption inhibition, as exemplified by agents such as raloxifene [7], bisphosphonates [8], and denosumab [9]. While these anti-resorptive therapies demonstrate fracture prevention efficacy, their long-term utility (> 5 years) remains clinically contentious [10]. Moreover, these agents are associated with significant adverse effects including atypical femoral fractures [10, 11] and bisphosphonate-related osteonecrosis of the jaw [12]. Importantly, their mechanism of action concomitantly suppresses bone formation alongside resorption inhibition [13]. Therefore, an increasing number of studies have focused on the development of drug targets for bone formation. The available anabolic armamentarium is limited to teriparatide (parathyroid hormone analog) [14], abaloparatide (PTH-related protein analog) [15], and romosozumab (sclerostin-neutralizing monoclonal antibody). Teriparatide and abaloparatide demonstrate vertebral bone mineral density improvement and hip fracture risk reduction [15], but their clinical application is restricted to 24 months owing to osteosarcoma risk with prolonged use [16]. Furthermore, discontinuation necessitates subsequent anti-resorptive therapy to maintain the acquired bone density [17]. Romosozumab presents similar limitations, with a 12-month treatment duration and potential cardiovascular complications [18]. These constraints underscore the urgent need to develop novel, safe, and durable anabolic agents with robust osteogenic potentials.

Notably, G protein-coupled receptors (GPCRs) represent the most prominent drug target family, with 34% of FDA-approved medications exerting therapeutic effects through GPCR modulation [19]. The GPCR superfamily comprises five principal classes: Rhodopsin-like (Class A), Secretin-like (Class B1), Adhesion (Class B2), Glutamate (Class C), and Frizzled (Class F) [20, 21]. Among these, Adhesion GPCRs remain the least characterized because of their structural complexity, and their precise regulatory functions in bone metabolism are poorly understood [22]. Our previous work identified G-protein coupled receptor 125 (Gpr125/Adgra3) as being highly expressed in osteoclasts, where it modulates differentiation via the AKT-NF-κB and MAPK signaling pathways [23]. This finding suggests that Gpr125 is an intriguing bidirectional modulator within the bone remodeling system; however, its potential function in osteogenesis, the pivotal anabolic counterpoint, remains to be elucidated.

In this study, we unravel this mystery by demonstrating that Gpr125 is a positive master regulator of osteoblast differentiation and bone formation. We report that Gpr125 is highly enriched in osteoblasts, with peak expression occurring on day 14 during pre-osteoblast differentiation. Knockdown of Gpr125 in pre-osteoblasts and bone marrow stromal cells (BMSCs) significantly impaired osteogenic differentiation and mineralization capacity, and markedly increased lipid droplet formation, mimicking the phenotype of postmenopausal and senile osteoporosis. Conversely, overexpression of Gpr125 at these stages reversed this phenotype. Combined with RNA-seq analysis, we successfully identified a downstream target gene of Gpr125, Gper1 (G protein-coupled estrogen receptor 1), and its involvement in the PI3K-AKT and Wnt/β-catenin signaling pathways. Subsequent rescue experiments confirmed that Gpr125 promotes bone formation and inhibits adipogenic differentiation via Gper1-mediated activation of the PI3K-AKT-β-catenin signaling cascade.

To further substantiate the pro-osteogenic role of Gpr125 in vivo, we engineered an adeno-associated virus (AAV) that overexpressed Gpr125 under the control of the Runx2 promoter (AAV9-Gpr125(mRunx2p-EGFP)). Injection of this virus into the calvaria and femurs of ovariectomized (OVX) and aging mice partially reversed this decrease in bone loss. Moreover, it restored the estrogen deficiency-induced decreased protein expression levels of Gper1, p-PI3K, p-AKT, and active β-catenin. Thus, our study moves beyond correlative observations to establish a definitive, hierarchically ordered signaling pathway through which Gpr125 promotes bone formation. These findings not only fundamentally advance our understanding of GPCR biology in skeletal homeostasis, but also suggest that the Gpr125-Gper1 axis is a promising and previously unexplored therapeutic target for developing next-generation anabolic treatments for osteoporosis.

Results

Gpr125 is abundantly expressed in osteoblasts both in vitro and in vivo.

To characterize the expression pattern of Gpr125 in bone tissue, total RNA was extracted from various tissues of one-month-old wild-type male mice and analyzed by quantitative reverse transcription polymerase chain reaction (qRT-PCR). Gpr125 expression was readily detectable in bone, with levels second only to those observed in the lung, visceral fat, subcutaneous fat, and kidney tissues (Fig. 1A). To further define the cellular localization of Gpr125 within osteoblasts, femoral sections from one-month-old wild-type male mice were co-stained with anti-Gpr125 and anti-Col1α1 antibodies, the latter serving as a marker for osteoblasts. Immunofluorescence analysis revealed clear co-localization of Gpr125 with Col1α1 on osteoblast membranes (Fig. 1B). Next, we assessed the temporal expression profile of Gpr125 during osteoblast differentiation. Primary calvarial cells (pre-osteoblasts) isolated from neonatal mice were subjected to osteogenic induction and proteins were harvested at various time points. Western blot analysis showed that Gpr125 expression peaked on day 14 of differentiation, followed by a decline (Fig. 1C, D). Based on these results, day 14 was identified as the critical time point and was selected for subsequent experiments.

Fig. 1.

Gpr125 is highly expressed in osteoblasts. (A) Expression levels of Gpr125 mRNA in various tissues of 1-month-old mice. (B) Immunofluorescence co-staining of Gpr125 (green) and Col1α1 (red) in femoral sections from neonatal mice. White arrows indicate sites of co-localization (yellow). (C) Protein expression levels of Gpr125 and Runx2 in murine calvarial cells cultured under osteogenic induction for 0, 4, 7, 14, and 21 days. (D) Quantification of protein expression shown in (C). NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Gpr125 knockdown in calvarial cells impairs osteogenesis and promotes adipogenesis

Following the lentiviral-mediated knockdown of Gpr125 in mouse calvarial cells (Figure S1), osteogenic induction was performed. On day 14, Alkaline Phosphatase (ALP) staining showed a pronounced decrease in ALP activity in Gpr125-knockdown cells compared to both the untreated control (mock) and the negative control (sh-NC) (Fig. 2A). Notably, an increase in vacuole formation (Fig. 2A, indicated by red arrows) was observed in the knockdown group, suggesting lipid droplet accumulation. This was confirmed by Oil Red O staining, which revealed abundant lipid droplets in Gpr125-deficient cells (Fig. 2B). Moreover, Alizarin Red S staining indicated a significant reduction in mineralization upon Gpr125 knockdown (Fig. 2C). To further investigate the mechanistic basis of these phenotypic changes, we examined the expression of the key lineage-specific markers. Western blot analysis confirmed efficient Gpr125 knockdown (88% reduction at the protein level), which led to markedly decreased levels of osteogenic transcription factors and markers, including Runx2 (76% reduction), OCN (78% reduction), Col1α1 (22% reduction), and OSX (85% reduction) (Fig. 2D, E). In contrast, the protein expression levels of the adipogenic regulators C/EBPα (3.4-fold increase) and PPARγ (2.8-fold increase) were significantly elevated (Fig. 2D, E). Consistent with these protein-level findings, qRT-PCR analysis confirmed successful Gpr125 silencing at the mRNA level and revealed corresponding transcriptional changes: downregulation of osteogenic genes (ATF4, ALP, OCN, Col1α1, and OSX) and upregulation of adipogenic genes (Fabp4, C/EBPα, and PPARγ) (Fig. 2F). Together, these results demonstrate that Gpr125 knockdown in pre-osteoblasts disrupts osteogenic differentiation and mineralization, while simultaneously enhancing adipogenic commitment.

Fig. 2.

Knockdown of Gpr125 in pre-osteoblasts attenuates osteogenic differentiation and promotes adipogenic differentiation. (A) Alkaline phosphatase (ALP) staining in calvarial-derived pre-osteoblasts from neonatal mice after 14 days of osteogenic induction, comparing Mock, sh-NC (non-targeting control), and Gpr125-knockdown (sh-Gpr125) groups. Red arrows indicate lipid-containing vacuoles. (B) Oil Red O staining of lipid droplets in Mock, sh-NC, and sh-Gpr125 groups following 14-day osteogenic induction. (C) Alizarin Red S staining of mineralized nodules in Mock, sh-NC, and sh-Gpr125 groups after 14 days of osteogenic induction. (D) Western blot analysis of Gpr125, Runx2, OCN, Col1α1, OSX, Pparγ, and C/ebpα protein expression in Mock, sh-NC, and sh-Gpr125 groups after 14 days of osteogenic induction. (E) Quantitative analysis of protein levels shown in (D). (F) mRNA expression levels of Gpr125, OSX, ATF4, OCN, Col1α1, ALP, Fabp4, Pparγ, and C/ebpα measured by qRT-PCR in sh-NC and sh-Gpr125 groups after 14 days of osteogenic induction. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Depletion of Gpr125 expression in BMSCs significantly inhibited osteogenesis and concomitantly potentiated adipogenesis

Since gene depletion at different stages of osteoblast differentiation can lead to distinct phenotypic outcomes [24], we investigated whether the effect of Gpr125 loss is consistent across differentiation stages. To this end, Gpr125 was knocked down at the earliest stage in BMSCs. After 14 days of osteogenic induction, ALP and Alizarin Red S staining revealed markedly reduced osteogenic differentiation and mineralization, respectively, in Gpr125-knockdown cells compared with controls. In contrast, Oil Red O staining revealed a pronounced increase in lipid accumulation (Fig. 3A). At the protein level, knockdown of approximately 40% Gpr125 significantly suppressed the expression of key osteogenic markers, including Runx2 (50% reduction), Col1α1 (75% reduction), OSX (95% reduction), and OCN (80% reduction) (Fig. 3B, C). Conversely, the adipogenic markers C/EBPα (4.7-fold increase) and PPARγ (3.3-fold increase) were markedly up-regulated (Fig. 3B, C). These findings were corroborated at the transcriptional level, which confirmed efficient Gpr125 knockdown and downregulated the expression of osteogenic genes (Runx2, Col1α1, OSX, OCN, and ATF4), along with the upregulation of adipogenic genes (Fabp4, C/ebpα, and PPARγ) (Fig. 3D). Together, these data indicate that Gpr125 depletion in BMSCs severely compromises osteogenic capacity and enhances adipogenic potential.

Fig. 3.

Gpr125 knockdown in BMSCs suppresses osteogenesis and promotes adipogenesis following 14 days of osteogenic induction. (A) Representative images of ALP staining, Alizarin Red S staining (mineralized nodules), and Oil Red O staining (lipid droplets) in BMSCs from Mock, sh-NC (non-targeting control), and sh-Gpr125 (Gpr125 knockdown) groups. (B) Western blot analysis of Gpr125, osteogenic markers (Runx2, Col1α1, OSX, OCN), and adipogenic markers (Pparγ, C/ebpα) in Mock, sh-NC, and sh-Gpr125 groups. (C) Quantitative analysis of protein expression levels from (B). (D) mRNA expression levels of Gpr125, osteogenic genes (Runx2, OSX, ATF4, OCN, Col1α1, ALP), and adipogenic genes (Fabp4, Pparγ, C/ebpα) measured by qRT-PCR in sh-NC and sh-Gpr125 groups. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Gpr125 overexpression in calvarial cells promotes osteogenesis and attenuates adipogenic protein expression

Given that Gpr125 knockdown in pre-osteoblasts impaired osteogenesis and enhanced adipogenesis, we investigated whether its overexpression would elicit the opposite effect. Calvarial cells were transduced with a Gpr125-overexpressing virus and subjected to osteogenic induction for 14 days. ALP and Alizarin Red S staining revealed a marked increase in osteogenic differentiation and mineralization, respectively, in the Gpr125-overexpressing group compared to the controls. In contrast, Oil Red O staining detected no significant lipid droplet formation in either group under these conditions (Fig. 4A). At the protein level, successful Gpr125 overexpression (2.7-fold increase) led to the upregulation of key osteogenic markers, including Runx2 (2.0-fold), Col1α1 (2.7-fold), OSX (3.8-fold), and OCN (2.4-fold). Conversely, the adipogenic markers PPARγ and C/EBPα were downregulated by 28 and 70%, respectively (Fig. 4B, C). qRT-PCR analysis confirmed efficient Gpr125 overexpression at the mRNA level and showed concomitant increases in the expression of osteogenic genes (Runx2, OSX, ATF4, OCN, Col1α1, and ALP). However, the mRNA levels of PPARγ and C/EBPα remained unchanged (Fig. 4D). These results indicate that overexpression of Gpr125 in calvarial cells significantly promotes osteogenic differentiation capability, while suppressing the expression of adipogenic genes at the protein level.

Fig. 4.

Overexpression of Gpr125 in pre-osteoblasts enhances osteogenic differentiation after 14 days of osteogenic induction. (A) Representative images of ALP staining, Alizarin Red S staining (mineralization), and Oil Red O staining (lipid droplets) in Mock, OE-NC (overexpression control), and OE-Gpr125 (Gpr125-overexpressing) groups. (B) Western blot analysis of Gpr125, osteogenic markers (Runx2, Col1α1, OSX, OCN), and adipogenic markers (Pparγ, C/ebpα) in Mock, OE-NC, and OE-Gpr125 groups. (C) Quantification of protein expression levels from (B). (D) mRNA expression levels of Gpr125, osteogenic genes (Runx2, OSX, ATF4, OCN, Col1α1, ALP), and adipogenic genes (Fabp4, Pparγ, C/ebpα) as determined by qRT-PCR in OE-NC and OE-Gpr125 groups. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Overexpression of Gpr125 in BMSCs significantly promoted osteogenesis while inhibiting adipogenesis

Given that Gpr125 overexpression at the pre-osteoblast stage enhances osteogenic differentiation while suppressing adipogenic potential, we examined whether similar phenotypic effects occur in BMSCs. To this end, Gpr125 was overexpressed in BMSCs, followed by osteogenic induction. After 14 days of differentiation, both ALP and Alizarin Red S staining were markedly stronger in the Gpr125-overexpressing group than in controls (Fig. 5A), indicating a pronounced increase in osteogenic differentiation capacity. In contrast, lipid droplet formation was not significantly different between groups (Fig. 5A). Western blot analysis confirmed an approximately 2-fold increase in Gpr125 in BMSCs (Fig. 5B, C), which was accompanied by the upregulation of key osteogenic markers, Runx2 (3.5-fold increase), OCN (2.4-fold increase), and Col1α1 (3.0-fold increase) (Fig. 5B, C). However, the protein levels of the adipogenic markers, PPARγ and C/EBPα, remained unchanged (Fig. 5B, C). At the mRNA level, a 2.5-fold increase in Gpr125 expression led to significant downregulation of PPARγ (30% reduction) and C/EBPα (85% reduction) transcripts (Fig. 5C). Conversely, mRNA levels of Runx2, OSX, ATF4, OCN, Col1α1, and ALP were substantially upregulated (Fig. 5C). These findings demonstrate that Gpr125 overexpression in BMSCs robustly enhances osteogenic differentiation capacity, while significantly suppressing adipogenic potential at the transcriptional level.

Fig. 5.

Overexpression of Gpr125 in BMSCs promotes osteogenic differentiation after 14 days of osteogenic induction. (A) Representative images of ALP staining, Alizarin Red S staining (mineralized nodules), and Oil Red O staining (lipid droplets) in Mock, OE-NC (overexpression negative control), and Gpr125-overexpressing (OE-Gpr125) groups of BMSCs. (B) Western blot analysis of Gpr125, Runx2, OCN, Col1α1, Pparγ, and C/ebpα protein expression in Mock, OE-NC, and OE-Gpr125 groups. (C) Quantitative analysis of protein expression levels shown in (B). (D) mRNA expression levels of Gpr125, Runx2, OSX, ATF4, OCN, Col1a1, ALP, Fabp4, Pparγ, and C/ebpα as determined by qRT-PCR in OE-NC and OE-Gpr125 groups. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Gper1 was targeted by Gpr125 in pre-osteoblasts and BMSCs

To elucidate the key mediators underlying impaired osteogenesis resulting from Gpr125 knockdown, we performed RNA sequencing (RNA-seq) on calvarial cells after 14 days of osteogenic induction. The volcano plot illustrates the transcriptomic changes upon Gpr125 depletion, with red, blue, and gray dots indicating significantly upregulated, downregulated, and unchanged genes, respectively (Fig. 6A). The top ten most upregulated and downregulated genes are shown in (Fig. 6A). We subsequently validated the top ten downregulated candidates by qRT-PCR. Consistent with the RNA-seq data, eight genes were significantly downregulated in Gpr125-knockdown cells compared to controls, whereas Pcdh17 and Pou3f4 remained unchanged. Notably, Gper1 expression was the most strongly suppressed, showing a 303-fold reduction (Fig. 6B). Given its pronounced downregulation, Gper1 was selected for further analyses. Western blotting of protein lysates collected after 14 days of osteogenic differentiation confirmed that Gpr125 knockdown significantly reduced Gper1 protein levels (Fig. 6C, D). Conversely, Gpr125 overexpression substantially increased Gper1 expression (Fig. 6E, F), indicating that Gpr125 positively regulated Gper1 expression in osteoblasts. Furthermore, immunofluorescence co-staining of differentiated BMSCs revealed the co-localization of Gpr125 and Gper1 on the cell membrane (Fig. 6G, white arrows), suggesting a potential membrane-localized interaction between these proteins during osteogenic differentiation.

Fig. 6.

Gpr125 promotes osteoblast differentiation by regulating the expression of its downstream gene Gper1. (A) Volcano plot displaying up- and down-regulated genes following Gpr125 knockdown in pre-osteoblasts after 14 days of osteogenic induction. (B) qRT-PCR validation of the top 10 most significantly down-regulated genes after Gpr125 knockdown. (C) Expression level of Gper1 upon successful knockdown of Gpr125 in pre-osteoblasts. (D) Quantification of the results shown in panel C. (E) Expression level of Gper1 upon successful overexpression of Gpr125 in pre-osteoblasts. (F) Quantification of the results shown in panel E. (G) Co-immunofluorescence staining of Gpr125 (Green) and Gper1 (Red) in BMSCs after 2 days of osteogenic induction (white arrows indicate co-localization of green and red signals). NS, not significant; *p < 0.05, **p < 0.01, ****p < 0.0001

Gpr125 regulates bone formation through the PI3K/AKT and Wnt/β-catenin signaling pathways

To explore the downstream signaling pathways modulated by Gpr125, we conducted a comprehensive analysis of RNA-seq data from Gpr125-knockdown calvarial cells. Gene Ontology (GO) analysis indicated that cell adhesion was the most significantly affected function upon Gpr125 knockdown (Fig. 7A), consistent with its known role as an adhesion molecule [25, 26]. Among the top ten biological processes most markedly suppressed, osteogenic mineralization ranked second (Fig. 7A), supporting a critical role for Gpr125 in osteoblast differentiation and mineralization. KEGG pathway analysis further identified the PI3K-AKT signaling pathway as the most significantly enriched among the top altered pathways (Fig. 7B), prompting its selection for further validation. Heatmap analysis demonstrated that Bone formation was markedly downregulated following Gpr125 knockdown, while Adipogenesis was significantly upregulated, which aligns with our above results. Given our prior reports linking Gpr125 to osteoclast differentiation via the ERK/MAPK pathway [23], we included this pathway in the heatmap. However, most ERK/MAPK pathway genes showed no significant changes upon Gpr125 knockdown (Fig. 7C). Based on the literature suggesting Gpr125 involvement in Wnt signaling [27, 28], we also analyzed the Wnt/β-catenin pathway and observed widespread downregulation of its components in Gpr125-deficient cells (Fig. 7C), leading us to investigate this pathway further. Next, we assessed the protein levels of the key signaling molecules. In pre-osteoblasts, Gpr125 knockdown significantly reduced the phosphorylation of PI3K and AKT (p-PI3K and p-AKT) (Fig. 7D, E), whereas Gpr125 overexpression increased their activation (Fig. 7F, G). Similarly, active β-catenin, a central mediator of Wnt signaling, decreased upon Gpr125 knockdown and increased upon its overexpression (Fig. 7H, I), confirming that both pathways are regulated by Gpr125. To determine whether these pathways operate independently or hierarchically, we treated Gpr125-overexpressing pre-osteoblasts with LY294002, a PI3K/AKT inhibitor [29, 30]. LY294002 abolished the Gpr125-induced increases in p-PI3K and p-AKT, and reduced the expression of the osteogenic marker Runx2 (Fig. 7J, K). Importantly, active β-catenin upregulation was also suppressed by LY294002, whereas Gpr125 protein levels remained unchanged (Fig. 7J, K). These results indicate that PI3K/AKT signaling acts upstream of Wnt/β-catenin activation in Gpr125-mediated osteogenic regulation, but not in reverse.

Fig. 7.

Gpr125 promotes osteoblast differentiation by regulating the PI3K/AKT and Wnt/β-catenin signaling pathways. (A) Gene Ontology (GO) analysis of the top 15 most significantly enriched biological processes among downregulated genes. (B) KEGG pathway analysis showing the top 10 most significantly enriched pathways in Gpr125-knockdown pre-osteoblasts. (C) Heatmap analysis depicting expression differences of genes related to bone formation, adipogenesis, ERK/MAPK, and Wnt/β-catenin signaling pathways between Gpr125-knockdown and negative control groups after 14 days of osteogenic induction. (D) Protein levels of p-PI3K and p-AKT in Gpr125-knockdown and negative control groups after 14 days of osteogenic induction. (E) Quantification of results shown in panel D. (F) Protein levels of p-PI3K and p-AKT in Gpr125-overexpression and negative control groups after 14 days of osteogenic induction. (G) Quantification of results shown in panel F. (H) Protein levels of active-β-catenin in Gpr125-knockdown and negative control groups, and in Gpr125-overexpression and negative control groups after 14 days of osteogenic induction. (I) Quantification of results shown in panel H. (J) Protein levels of p-PI3K, p-AKT, active-β-catenin, Gpr125, and Runx2 in OE-NC, OE-NC + DMSO, and OE-Gpr125 + LY294002 groups after treatment with LY294002 in Gpr125-overexpressing pre-osteoblasts. (K) Quantification of results shown in panel J. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Gper1 functions upstream of Gper1 and β-Catenin without Feedback Regulation

Although Gper1 was identified as a downstream target of Gpr125, it remains unclear whether Gper1 exerts retrograde regulation of Gpr125 expression. To address this, we overexpressed Gper1 in Gpr125-knockdown BMSCs. After 14 days of osteogenic induction, ALP staining showed that Gper1 overexpression rescued the impaired ALP activity resulting from Gpr125 knockdown (Fig. 8A). To simultaneously evaluate adipogenic changes under osteogenic conditions, we performed sequential ALP and Oil Red O staining of the same cultures. This revealed that enhanced lipid droplet formation upon Gpr125 knockdown was markedly suppressed by Gper1 overexpression (Fig. 8A). At the protein level, Gper1 overexpression in Gpr125-deficient cells restored the expression of osteogenic markers (Runx2 and OCN) and key signaling molecules (p-PI3K, p-AKT, and active β-catenin). Notably, however, Gpr125 protein levels were not recovered and were further reduced (Fig. 8B, C), indicating that Gper1 does not regulate Gpr125 but can activate the PI3K-AKT-β-catenin axis independently. Next, we investigated whether β-catenin exerts retrograde control on Gpr125 or the Gper1/PI3K-AKT axis. Treatment of Gpr125-knockdown BMSCs with Wnt3a-conditioned medium (Wnt3a CM) [31], a Wnt pathway activator, successfully elevated active β-catenin levels above baseline levels (Fig. 8D, E). However, this did not restore Gpr125 expression, nor did it rescue the reduced levels of p-PI3K, p-AKT, or Runx2 (Fig. 8D, E). Interestingly, OCN expression was partially recovered. Functionally, ALP and Oil-Red staining showed that active β-catenin induction reversed the osteogenic deficit and abolished the enhanced adipogenesis seen in Gpr125-knockdown cells (Fig. 8F). Together, these data demonstrate that β-catenin acts downstream of the Gpr125/Gper1/PI3K-AKT pathway and does not regulate upstream components.

Fig. 8.

Gpr125 promotes osteogenic differentiation and inhibits adipogenic differentiation via Gper1-mediated activation of the PI3K–AKT–β-catenin signaling pathway. (A) After 14 days of osteogenic induction, ALP staining was performed followed by Oil Red O staining on the same dish of cells in the sh-NC, sh-Gpr125 + OE-NC, and sh-Gpr125 + OE-Gper1 groups. (B) Protein expression levels of Gper1, Gpr125, OCN, Runx2, active-β-catenin, p-PI3K, and p-AKT in the sh-NC, sh-Gpr125 + OE-NC, and sh-Gpr125 + OE-Gper1 groups. (C) Quantitative analysis of the results shown in panel B. (D) Protein expression levels of active-β-catenin, Gpr125, Runx2, OCN, p-PI3K, and p-AKT in the sh-NC, sh-Gpr125, and sh-Gpr125 + Wnt3a CM groups. (E) Quantitative analysis of the results shown in panel D. (F) After 14 days of osteogenic induction, ALP staining and Oil Red O staining were sequentially performed on cells from the sh-NC, sh-Gpr125, and sh-Gpr125 + Wnt3a CM groups. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Gpr125 overexpression partially rescues bone loss in ovariectomized mice.

Our in vitro findings demonstrate that Gpr125 promotes osteogenesis via Gper1-mediated activation of the PI3K/AKT/β-catenin signaling axis. To evaluate whether this pathway functions similarly in vivo, we established an ovariectomized (OVX) mouse model using 2-month-old female mice subjected to bilateral ovariectomy. An adeno-associated virus (AAV) expressing Gpr125 under the Runx2 promoter was locally injected into calvariae and femurs to target osteoprogenitor cells. Bone samples were collected for micro-computed tomography (micro-CT) and histological analysis (Fig. 9A). AAV localization was confirmed by GFP fluorescence, which was restricted to the calvariae and femurs, with a negligible signal in the off-target organs (Fig. 9B), indicating site-specific overexpression. confirmed the targeted overexpression at these skeletal sites. Micro-CT analysis of the calvaria revealed a pronounced reduction in bone mass in OVX mice injected with the control virus (OVX + AAV-NC) compared to that in the sham-operated group (Sham), as indicated by the increased purple coloration. Conversely, this bone loss phenotype was markedly attenuated in OVX mice that received Gpr125-overexpressing virus (OVX + AAV-Gpr125) (Fig. 9C). Quantification revealed significantly decreased bone volume fraction (BV/TV), cortical thickness (Ct.Th), and trabecular thickness (Tb.Th), as well as increased bone surface area per unit volume (BS/BV) in the OVX + AAV-NC group compared to the Sham group (Fig. 9D). In contrast, the OVX + AAV-Gpr125 group exhibited a significantly higher BV/TV than the OVX + AAV-NC mice, although other microstructural parameters did not differ significantly (Fig. 9D). Femoral micro-CT analysis similarly indicated severe bone loss in OVX + AAV-NC mice relative to sham mice (Fig. 9E). Quantitative assessment showed significant reductions in bone mineral density (BMD), BV/TV, Ct.Th, Tb.Th, and trabecular number (Tb.N) and increased BS/BV and trabecular separation (Tb.Sp) (Fig. 9F). Gpr125 overexpression in OVX mice partially reversed these changes, restoring BMD, BV/TV, Tb.Th, and Tb.N to sham levels, and normalizing BS/BV and Tb.Sp (Fig. 9F). To assess the signaling activity in bone tissue, we analyzed femoral protein lysates. Western blotting showed that OVX + AAV-NC mice had significantly lower levels of Gper1, p-PI3K, p-AKT, and active β-catenin than the sham controls. These reductions were effectively rescued by Gpr125 overexpression (Figure S2A, B). Hematoxylin and eosin (H&E) staining of decalcified femoral sections revealed trabecular bone loss and substantial lipid droplet accumulation in the marrow cavity of OVX + AAV-NC mice. Both phenotypes were markedly improved in the OVX + AAV-Gpr125 group (Fig. 9G). Collectively, these in vivo results substantiate the role of Gpr125 in promoting bone formation and mitigating estrogen deficiency-induced bone loss.

Fig. 9.

Gpr125 overexpression rescues low bone mass in the calvaria and femur of ovariectomized mice with postmenopausal osteoporosis. (A) Schematic diagram of adeno-associated virus carrying Gpr125 under the Runx2 promoter injected into the calvaria and femur of bilateral ovariectomized (OVX) mice. (B) GFP fluorescence expression in various tissues and organs after injection of Gpr125 adeno-associated virus. (C) Changes in bone mass in the calvaria of mice following injection of Gpr125 adeno-associated virus. (D) Quantitative micro-CT analysis of bone microarchitectural parameters for the bone mass shown in panel C. (E) Changes in bone mass in the femur of mice after injection of Gpr125 adeno-associated virus. (F) Quantitative micro-CT analysis of bone microarchitectural parameters for the bone mass shown in panel E. (G) Representative H&E-stained sections of the femur from mice in the Sham, OVX + AAV9-NC, and OVX + AAV9-Gpr125 groups. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Overexpression of Gpr125 reversed bone loss in aged mice

While these findings establish that Gpr125 overexpression can rescue postmenopausal osteoporosis, its potential efficacy against age-related osteoporosis remains unclear. To address this, AAV-Gpr125 was injected into the calvariae and femora of 23-month-old male mice and harvested one-month post-injection. Micro-CT analysis revealed that 23-month-old mice exhibited significantly reduced bone mass compared to 3-month-old mice in both the calvarial and femoral regions (Fig. 10), indicating the successful establishment of a senile osteoporosis mouse model. Following the injection of Gpr125 adeno-associated virus into the calvariae of aged mice, the overexpression group (AAV-Gpr125) showed a 1.4-fold increase in BMD, 1.2-fold increase in BV/TV, 2.3-fold increase in Tb.Th, and 1.3-fold increase in Tb.N compared to the control group (AAV-NC) (Fig. 10A, B). Similarly, injection of Gpr125 adeno-associated virus into the femora of aged mice resulted in a 1.1-fold increase in BMD, a 2.2-fold increase in BV/TV, a 1.8-fold increase in Tb.Th, a 2.7-fold increase in Tb.N, and an approximately 20% decrease in Tb.Sp (Fig. 10A, B). These findings demonstrate that Gpr125 overexpression partially reversed age-related bone loss in mice.

Fig. 10.

Gpr125 overexpression reverses low bone mass in aged mice. (A) Micro-CT images of calvarial bone from 23-month-old male mice injected with Gpr125-overexpressing adeno-associated virus. (B) Quantitative analysis of bone microarchitectural parameters corresponding to (A). (C) Micro-CT images of femoral bone from 23-month-old male mice injected with Gpr125-overexpressing adeno-associated virus. (D) Quantitative analysis of bone microarchitectural parameters corresponding to (C). NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

A working model wherein Gpr125 activation promotes bone formation and suppresses lipid accumulation

Mechanistically, the extracellular N-terminus of Gpr125 features key structural domains, including leucine-rich repeats (LRRs), immunoglobulin (Ig) domains, hormone-binding domains (HBD), and GPCR autoproteolysis-inducing (GAIN) domains. Under basal conditions, the tethered agonist peptide is constrained within the GAIN domain. Upon induction of osteogenic differentiation, cis-autoproteolytic cleavage occurs, releasing the agonist and leading to the self-activation of the receptor. This event triggers the upregulation of Gper1, which in turn activates the PI3K/AKT signaling pathway and stabilizes β-catenin, collectively driving osteogenesis while inhibiting adipogenic lipid droplet formation (Fig. 11).

Fig. 11.

A working model whereby Gpr125 promotes bone formation and inhibits lipid droplet formation. Schematic diagram illustrating the proposed mechanism by which Gpr125 promotes bone formation and suppresses adipogenic differentiation

Discussion

G protein-coupled receptors (GPCRs), with over 800 members in humans, serve as critical sensors of extracellular stimuli through their characteristic seven-transmembrane domains, facilitating intracellular signal transduction [20, 32, 33]. Structurally, the GPCR superfamily is divided into five major classes: rhodopsin-like (Class A), secretin receptor (Class B1), adhesion (Class B2), glutamate (Class C), and frizzled (Class F) [20, 21]. Notably, many key regulators of bone remodeling, such as RANKL, PTH, estrogen, and Wnt ligands function through GPCRs, underscoring their importance as therapeutic targets for bone metabolic disorders such as osteoporosis [22, 34]. Among these, adhesion GPCRs (aGPCRs) remain among the least understood classes because of their structural complexity, and their specific roles in bone metabolism remain largely unexplored [22]. Among the 33 members, aGPCRs represent the second largest GPCR class and are broadly expressed in ectoderm-, mesoderm-, and endoderm-derived tissues, where they regulate diverse processes including development, reproduction, neurosynapse formation, angiogenesis, immunity, and tumorigenesis [21, 35]. Although initially recognized for mediating cell–cell adhesion [36], recent studies have revealed their involvement in sensing external cues and transducing signals through multiple pathways [37]. In a previous study, we identified Gpr125, a prominent aGPCR that is highly expressed in monocytes/macrophages and osteoclasts, and demonstrated its role in osteoclast differentiation via AKT–NF-κB and MAPK signaling [23]. Given that bone homeostasis depends on the balance between osteoblast-driven bone formation and osteoclast-mediated bone resorption [38, 39], we investigated whether Gpr125 also influences osteogenesis. Our study unveils a previously unrecognized anabolic role of the adhesion GPCR Gpr125 in bone formation and delineates a complete signaling axis, from receptor to downstream effector, that governs osteoblast-adipocyte fate commitment. The discovery that Gpr125 activation promotes osteogenesis through a novel Gper1-PI3K/AKT-β-catenin cascade not only resolves a key gap in GPCR biology but also opens a new therapeutic avenue for treating osteoporosis by targeting a pro-formation pathway.

To comprehensively assess the role of Gpr125 across the osteogenic lineage, we employed two primary cell models: mouse calvarial cells, representing the pre-osteoblast stage, and BMSCs, representing an early multipotent mesenchymal stage. This dual-cell approach is supported by our prior observation that genetic perturbation at distinct differentiation stages can yield divergent phenotypic outcomes [24]. Consistent with this premise, Gpr125 knockdown, albeit with different efficiencies (88% in pre-osteoblasts vs. 40% in BMSCs), differentially impaired osteogenic markers. In preosteoblasts, knockdown reduced Runx2, OCN, Col1α1, and Osterix expression by 76, 78, 22, and 85%, respectively. Notably, a more modest knockdown in BMSCs resulted in profound reductions of 50%, 80%, 75%, and 95% in the same markers, indicating greater sensitivity to Gpr125 levels at this primitive stage. This heightened impact was corroborated by the more robust enhancement of osteogenesis upon Gpr125 overexpression in BMSCs compared to pre-osteoblasts. Collectively, these in vitro data demonstrate that Gpr125 exerts a more potent effect on the osteogenic commitment of BMSCs, which gives rise to long bones, than on the maturation of calvaria-derived pre-osteoblasts. We translated these findings into pathophysiologically relevant in vivo models. In a postmenopausal osteoporosis (PMOP) model characterized by high bone turnover with net bone loss [40, 41], we selectively targeted osteoblasts using an adeno-associated virus (AAV) that drives Gpr125 expression under the Runx2 promoter. This strategy isolated the anabolic role of Gpr125 from its known effects on osteoclasts. AAV-mediated overexpression of Gpr125 led to a modest improvement in BV/TV in calvariae, but elicited a more robust anabolic response in femora, significantly enhancing BMD, BV/TV, Tb.Th, and Tb.N. The observed bone mass recovery was mechanistically linked to activated osteogenesis, as evidenced by the elevated Runx2 expression in Gpr125-overexpressing mice. This site-specific efficacy was further corroborated in a senile osteoporosis model characterized by low bone turnover [42, 43], where Gpr125 overexpression elicited a superior restorative effect in long bones (femora) than in flat bones (calvariae). Consistent in vitro and in vivo evidence strongly establishes Gpr125 is a significant pro-osteogenic factor with a predominant role in the physiology of long bones.

Knockdown of Gpr125 in both pre-osteoblasts and BMSCs not only suppressed osteogenic differentiation but also concurrently promoted adipogenesis, as evidenced by enhanced lipid droplet accumulation. This shift in cell fate commitment mirrors the hallmark pathophysiological features of postmenopausal and senile osteoporosis, namely reduced bone mass coupled with increased marrow adipose tissue (MAT) [24, 44], underscoring the physiological relevance of Gpr125 in maintaining skeletal metabolic balance. To elucidate the mechanism underlying the potent pro-osteogenic function of Gpr125, we performed transcriptomic profiling followed by Gene Ontology (GO) analysis. The results indicated that Gpr125 knockdown led to the most pronounced downregulation of biological processes related to cell adhesion and osteoblast mineralization. This suggests that, beyond its canonical role in cell adhesion, Gpr125, an adhesion-family GPCR, plays a more specialized role in driving osteogenic differentiation. Through integrated analysis of volcano plots and qRT-PCR validation, we identified Gper1 as a key downstream target of Gpr125. Subsequent protein-level assays confirmed that Gpr125 regulates Gper1 expression in a unidirectional manner. KEGG pathway analysis further highlighted the PI3K-AKT signaling pathway as the most significantly enriched pathway upon Gpr125 perturbation, a finding corroborated by western blot analysis. Notably, although prior studies have implicated Gpr125 in non-canonical Wnt/PCP signaling during zebrafish gastrulation [45, 46] and as a suppressor of Wnt/β-catenin signaling in colorectal cancer [47], its role in osteogenic Wnt signaling remains unexplored. Given the pivotal role of the Wnt/β-catenin signaling pathway in promoting bone formation and development [48, 49], these findings suggest an important connection between Gpr125 and the Wnt/β-catenin pathway. While the Wnt/β-catenin pathway was not ranked among the top enriched pathways in our KEGG analysis, heatmap visualization revealed a broad downregulation of genes associated with this pathway following Gpr125 knockdown. Subsequent experimental validation confirmed that Gpr125 indeed regulates the active form of β-catenin (dephosphorylated at Ser37/Thr41). Together, these findings suggest that Gpr125 is an upstream regulator of three critical signaling components: Gper1, PI3K-AKT, and Wnt/β-catenin. Having established these connections, we next sought to delineate the functional hierarchy of these signaling entities.

Gper1 (also known as Gpr30) is a GPCR broadly expressed in bone cells, including osteoblasts, osteoclasts, osteocytes, and chondrocytes [50]. Studies using Gpr30 knockout mice have established their importance in skeletal development [51, 52], and they are known to mediate rapid, non-genomic signaling in response to estradiol (E2), complementing the classical genomic actions of ERα and ERβ in bone protection [51, 53]. In osteogenic differentiation, Gper1 has been reported to promotes osteoblast maturation via the PI3K/AKT pathway [54–56]. In the present study, we found that Gpr125 regulates PI3K/AKT signaling both directly and through Gper1. Specifically, Gper1 overexpression rescued the impaired phosphorylation of PI3K and AKT resulting from Gpr125 knockdown. Conversely, PI3K inhibitor treatment abolished the enhanced p-PI3K and p-AKT levels induced by Gpr125 overexpression, without affecting Gpr125 or Gper1 expression, indicating that PI3K/AKT acts downstream of Gper1 and does not exert feedback regulation on either GPCR. This places Gper1 upstream of PI3K/AKT in the osteogenic signaling cascade. The PI3K/AKT pathway is well recognized as a key promoter of bone formation [57, 58]. Xi et al. demonstrated that the protein levels of phosphorylated PI3K and phosphorylated AKT were significantly reduced in the bone tissue of osteoporotic rats and that the PI3K/AKT pathway might inhibit osteoporosis by promoting osteoblast proliferation and bone formation [59]. In OVX mice, activation of the PI3K/AKT pathway promotes bone formation and prevents bone resorption [60, 61], consistent with our findings. Our results showed that the PI3K/AKT pathway promotes osteogenic differentiation and mineralization, and ovariectomy (OVX) downregulates the expression of p-PI3K and p-AKT in the femur. More importantly, Gpr125 overexpression significantly increased the expression of p-PI3K and p-AKT in the femora of OVX mice, thereby rescuing estrogen withdrawal-induced bone loss. Both PI3K/AKT and Wnt/β-catenin are key signaling pathways involved in bone formation. Next, we investigated the relationship between PI3K/AKT and Wnt/β-catenin signaling, another essential osteogenic pathway. Previous work by Dong et al. demonstrated that in osteoblastic cells and fracture healing models, PI3K/AKT signaling promotes β-catenin activation via phosphorylation of GSK-3β (Ser9) and β-catenin (Ser552), thereby stabilizing β-catenin and facilitating its nuclear translocation [62]. In our system, PI3K inhibition abolished the upregulation of active β-catenin (dephosphorylated at Ser37/Thr41) induced by Gpr125 overexpression. Conversely, although Wnt3a-conditioned medium restored active β-catenin expression in Gpr125-knockdown cells, it failed to rescue the suppression of p-PI3K and p-AKT levels. These results unequivocally demonstrate that in Gpr125-mediated osteogenesis, PI3K/AKT functions upstream of Wnt/β-catenin, and both pathways operate downstream of Gpr125.

Skeletal aging is pathologically characterized by a decline in bone mass coupled with the expansion of marrow adipose tissue (MAT) [63], a hallmark feature shared by both postmenopausal and senile osteoporosis [41, 44]. Numerous studies have shown that the PI3K/AKT signaling pathway plays a major role in promoting osteogenesis and inhibiting adipogenesis, thereby maintaining bone-fat metabolic balance [64, 65]. Similarly, many studies have elucidated the critical role of the Wnt/β-catenin signaling pathway in promoting bone formation and inhibiting adipogenic differentiation, thus delaying skeletal aging [66–68]. In line with these established roles, we found that Gpr125 knockdown in osteoblastic cells not only suppressed osteogenic mineralization and marker expression (e.g., Runx2 and Osterix) but also enhanced lipid accumulation and adipogenic gene expression (C/ebpα and Pparγ). These changes were accompanied by reduced activation of both PI3K/AKT and β-catenin signaling. Conversely, Gpr125 overexpression enhanced osteogenic outcomes and pathway activity while effectively suppressing adipogenic differentiation. These in vitro results phenocopied the osteopenic and adipocyte-rich marrow environment that is typical of skeletal aging. Importantly, the therapeutic relevance of these findings has been confirmed in vivo. AAV-mediated overexpression of Gpr125 in osteoporotic mouse models not only ameliorated bone loss, but also reduced MAT accumulation, concurrently restoring the impaired activity of the PI3K/AKT/β-catenin signaling axis in bone tissue. Taken together, our results establish Gpr125 as a key modulator of skeletal aging and demonstrate that its protective effects are mediated, at least in part, through the coordinated activation of osteogenesis-supportive PI3K/AKT and Wnt/β-catenin signaling.

Gpr125 (Adgra3), an adhesion GPCR (aGPCR), and an orphan receptor feature an N-terminal extracellular architecture comprising leucine-rich repeats (LRR), an immunoglobulin (Ig) domain, a hormone-binding domain (HBD), and a GAIN domain (26). Like many aGPCRs, Gpr125 is predicted to undergo autoproteolysis at a conserved GPCR proteolysis site (GPS) within its GAIN domain [26]. Unlike classical GPCRs, which rely on ligand binding for activation, aGPCRs are increasingly recognized to utilize a tethered agonist mechanism intrinsic to the GAIN domain that is exposed upon cleavage and can trigger receptor signaling in a ligand-independent manner [21, 69]. This model suggests that Gpr125 may function through constitutive or context-dependent self-activation. Building on this paradigm and supported by our experimental data, we propose that Gpr125 activation, potentially mediated by its tethered agonist, regulates Gper1 expression, thereby initiating the downstream PI3K/AKT/β-catenin signaling cascade that drives osteoblast differentiation, suppresses adipogenesis, and confers protection against osteoporosis and skeletal aging.

Experimental section

Ethics approval statement

The wild-type (C57BL/6) mice used in this study were purchased from Hunan Slek Jingda Laboratory Animal Co., Ltd. All animal experiments were conducted in accordance with guidelines approved by the Institutional Animal Care and Use Committee of the Third Xiangya Hospital of Central South University.

Primary cell culture and osteogenesis induction

Calvariae were harvested from wild-type neonatal mice (postnatal day 1–3) and digested with type I collagenase (SCR103; Sigma-Aldrich) to isolate the primary cells. The cells were then cultured in α-MEM low-glucose medium (HyClone, USA) supplemented with 10% fetal bovine serum (A5669701, Gibco, Australia) and 1% penicillin/streptomycin (C0222, Beyotime, China), and maintained at 37 °C in a 5% CO₂ humidified incubator. Wild-type male mice (2–3 months old) were euthanized by cervical dislocation. Femora and tibiae were aseptically isolated, and bone marrow stromal cells (BMSCs) were flushed out with the culture medium. The cells were subjected to adherent cultured in α-MEM low-glucose medium (HyClone, USA) containing 15% fetal bovine serum (A5669701, Gibco, Australia) and 1% penicillin/streptomycin (C0222, Beyotime, China) at 37 °C under 5% CO₂. Cells from both sources at passage 3 were used for osteogenic induction. The induction medium consisted of a basal culture medium supplemented with 50 µg/mL ascorbic acid (1043003, Sigma-Aldrich), 10 mM β-glycerophosphate (G9422, Sigma-Aldrich), and 10 nM dexamethasone (D4902, Sigma-Aldrich). To suppress the PI3K/AKT axis, cells were pre-incubated for 1 h with 10 µM LY294002 (S1105, Selleck Chemicals, USA). To activate Wnt/β-catenin signal, wnt3a conditional medium (wnt3a CM) is prepared from L Wnt-3 A (ATCC CRL-2647, USA) cells according to the manufacturer’s instructions.

Gene knockdown and overexpression

Lentiviral vectors expressing shRNA targeting the mouse Gpr125/Adgra3 gene (target sequence: 5′-GCTGGGTAAAGGAGAGAAACA-3′) or a negative control shRNA (5′-TTCTCCGAACGTGTCACGT‐3′) were constructed using the GV493 backbone (pFU-GW-016; Genechem, China) with BsmBI sites. All constructs were verified by Sanger sequencing. For overexpression, full-length coding sequences of mouse Gpr125/Adgra3 (NM_133911.1) and Gper1 (NM_029771.3) were cloned from a cDNA library (Genechem) using gene-specific primers: Gpr125: Forward, 5′‐GATCTATTTCCGGTGAATTCCGCCACCATGGAGCCGCCGCCGCCGCTGCTGCTGCTG‐3′, reverse: 5′-TCCTTGTAGTCCATGGATCCCACAGTTGTTTCGTGTTTCCATAACCC‐3′; Gper1: Forward: 5′‐GATCTATTTCCGGTGAATTCCGCCACCATGGATGCGACTACTCCAGC‐3′ Reverse: 5′-TCCTTGTAGTCCATGGATCCCACAGCACTGCTGAACCTGACCTC-3′. The fragments were inserted into the lentiviral vectors GV640 (for Gpr125) and GV828 (for Gper1) (both from Genechem) via EcoRI and BamHI restriction sites using an In-Fusion HD cloning kit. The positive clones were sequenced. Lentiviral particles were produced by co-transfecting 293T cells with the transfer and packaging plasmids psPAX2 and pMD2.G, respectively, using Lipofectamine 2000 (Invitrogen). Viral supernatants were collected 72 h post-transfection, centrifuged at 2000 × g for 10 min, filtered through 0.45 μm filters, and titrated by flow cytometry based on GFP or mCherry expression. Viral stocks with titers of approximately 1 × 10⁹ TU/mL were aliquoted and stored at − 80 °C. Calvarial osteogenic cells and bone marrow stromal cells (BMSCs) were infected with the respective lentiviruses to knock down or overexpress target genes, as required for subsequent experiments.

RNA isolation and quantitative real-time PCR (qRT‐PCR)

Total RNA was isolated from calvarial cells, BMSCs, and mouse tissues using TRIzol reagent (Thermo Fisher Scientific, USA) following the manufacturer’s protocol. cDNA synthesis was performed with 1 µg of total RNA using a PrimeScript RT Reagent Kit with gDNA Eraser (RR047A, TaKaRa, Japan) and random hexamer primers. Quantitative real-time PCR (qRT-PCR) was performed using a SYBR Premix Ex Taq II kit (RR820A, TaKaRa, Japan) on a Bio-Rad CFX Connect Real-Time PCR Detection System (Bio-Rad, USA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an endogenous reference gene for normalization. All the primer sequences used for qRT-PCR are listed in Table S1 (Supporting Information).

Western blot

Total protein was extracted using RIPA buffer (10 mM Tris-HCl, 1% Nonidet P-40, 0.1% SDS, 150 mM NaCl, and 1 mM EDTA [pH 7.5]) supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF), 5 µM leupeptin, and 10 µM aprotinin. Protein concentrations were quantified using a bicinchoninic acid (BCA) assay (Thermo Fisher Scientific, USA) with bovine serum albumin (BSA) as the standard. Total cell lysates were separated on 8% or 10% (w/v) sodium dodecyl sulfate-polyacrylamide gels and subsequently transferred onto Immobilon PVDF membranes (EMD Millipore, Billerica, MA, USA). Membranes were blocked with 5% non-fat dry milk in Tris-buffered saline containing 0.1% Tween 20 (TBST; 50 mM Tris, pH 7.6, 150 mM NaCl, 0.1% Tween 20) and incubated overnight at 4 °C with primary antibodies diluted in TBST containing 3% non-fat dry milk. The following primary antibodies were used: anti-Gpr125 (1:500, PA5-120377, Invitrogen, USA), anti-Runx2 (1:1000, 20700-1-AP, Proteintech, China), anti-Col1α1 (1:1000, 67288-1-Ig, Proteintech, China), anti-Osx (1:1000, ab209484, Abcam, UK), anti-OCN (1:1000, DF12303, Affinity Biosciences, USA), anti-C/ebpα (1:1000, AF6333, Affinity Biosciences, USA), anti-PPARγ (1:1000, AF6284, Affinity Biosciences, USA), anti-Gper1 (1:1000, DF2737, Affinity Biosciences, USA), anti-p-AKT (1:1000, #4060, Cell Signaling Technology, USA), anti-AKT (1:1000, #2920, Cell Signaling Technology, USA), anti-p-PI3K (1:1000, #4228, Cell Signaling Technology, USA), anti-PI3K (1:1000, #4257, Cell Signaling Technology, USA), anti-Active-β-catenin (1:1000, 05-665, Millipore, USA), anti-β-catenin (1:1000, #8480, Cell Signaling Technology, USA), and anti-GAPDH (1:4000, 60004-1-Ig, Proteintech, China). After incubation with the primary antibody, the membranes were probed with HRP-conjugated secondary antibodies: goat anti-rabbit IgG (H + L) (1:10,000, bs-0295G, Bioss, China) or goat anti-mouse IgG (H + L) (1:10,000, bs-0296G, Bioss, China). A protein marker (26616, Thermo Fisher Scientific, USA) was used according to the manufacturer’s instructions. Protein bands were visualized and quantified using Quantity One software (Bio-Rad, USA) and further analysis was performed using ImageJ software (NIH, USA). Original uncropped blot images are provided in Figure S3 (Supporting Information).

Immunofluorescence (IF) staining

Paraffin tissue sections were deparaffinized and antigen retrieval was achieved by heat treatment using a commercial reagent (Abcam AB970). Mouse femurs were analyzed by immunofluorescence using the following primary antibodies: rabbit-anti-Gpr125 (1:300, 11912-1-AP, Proteintech, China), mouse-anti-Col1α1 (1:300, 67288-1-Ig, Proteintech, China). The secondary antibodies used were goat anti-rabbit IgG (H + L) -FITC (1:200, SA00003-2, Proteintech, China) and goat anti-mouse IgG (H + L)-FITC (1:200, SA00003-1, Proteintech, China). Slides were counterstained with DAPI (C1002, Beyotime, China) for 10 min at room temperature. To obtain adherent cells, BMSCs were seeded onto glass coverslips and cultured for 48 h. Co-staining of Gpr125 and Gper1 was performed using a three-color fluorescence kit (RCF0086-23R-Fr, Huilan Biotech, Shanghai, China) based on tyramide signal amplification (TSA) technology, following the manufacturer’s instructions. After fixation with 4% paraformaldehyde (PFA) for 15 min, cells were blocked with 1% bovine serum albumin (BSA) (ST2254, Beyotime, China) in phosphate-buffered saline (PBS) for 30 min at room temperature. Subsequently, the cells were incubated with an anti-Gpr125 primary antibody (1:200, 11912-1-AP, Proteintech, China) overnight at 4 °C. To perform TSA-based staining for Gpr125 and Gper1, cells initially stained for Gpr125 were treated with a clearing buffer to denature and remove bound antibodies. The cells were then washed thrice with PBS containing Tween 20 (PBST) at room temperature to eliminate residual SDS. The samples were then incubated with an anti-Gper1 primary antibody (1:200, DF2737, Affinity Biosciences, USA) and stained with the corresponding tyramide reagent. All steps were carried out in the dark to prevent photobleaching. The cell nuclei were counterstained with DAPI (C1002, Beyotime, China) for 10 min at room temperature. Finally, the samples were imaged under a fluorescence microscope (Mshot, China).

Alkaline phosphatase (ALP), Oil Red O, and Alizarin Red S (ARS) Staining

The calvarial cells and BMSCs were fixed in 4% paraformaldehyde for 30 min. ALP staining was performed using a BCIP/NBT Alkaline Phosphatase Color Development Kit (C3206, Beyotime, China) according to the manufacturer’s instructions. Lipid accumulation was assessed using an Oil Red O Staining Kit (C0158S, Beyotime, China) according to the manufacturer’s protocol. Mineralization was evaluated by staining fixed cells with 2% Alizarin Red S solution (pH 4.2; C0138, Beyotime, China) following the supplier’s guidelines.

RNA-sequencing analysis

Total mRNA was extracted from osteoblasts cultured for 14 days in osteogenic differentiation medium using TRIzol reagent (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. The extracted RNA was sent to GENESEED (Guangzhou, China) for quality assessment using an Agilent 2100 Bioanalyzer. Sequencing libraries were constructed using the GENESEED RNA-seq Library Prep Kit. Reads were aligned to the GENCODE reference genomes (human-hg19, mouse-mm10) using Bowtie2 (version 2.1.0) and gene expression levels were quantified using RSEM v1.2.28. Normalization was performed using the TMM (trimmed mean of M-values) method. Differential gene expression analyses were conducted using the edgeR package. Genes with a p-value < 0.05 and | log₂ (fold change) | > 0.585 (corresponding to 1.5-fold) were considered differentially expressed. A volcano plot was generated based on log₂ (fold change) and log10 (p-value). Gene Ontology (GO) enrichment analysis was performed using the DAVID online tool (https://david.ncifcrf.gov/). The most significantly enriched GO terms among the downregulated genes were selected based on p-values and enrichment scores, and the number of downregulated genes in each category was determined. Pathway analysis was performed using the KEGG database. The top enriched KEGG pathways were selected according to p-values and enrichment scores. An online heatmap tool (http://www.heatmapper.ca/expression/) was used to visualize the expression patterns of DEGs identified in the KEGG pathways. Original data are provided in Supplementary Material.

Construction of AAV9-Gpr125

Plasmid construction

The adeno-associated virus vector expressing small activating RNA (saRNA) of the Gpr125/Adgra3 gene (GGAAAGAGCCTGAAACCAATC) and negative control (TTCTCCGAACGTGTCACGT) were synthesized and cloned into the mRUNX2p-EGFP-MIR155(NC)-WPRE-SV40 polyA vector with NheI and XhoI sites (purchased from Shanghai Genechem Co., Ltd.), and the recombinant vector was detected by DNA sequencing.

Adeno-Associated Virus Production

The recombinant AAV plasmid was co-transfected into HEK293T cells with the pHelper and pAAV9-RC plasmids (AAV9 serotype packaging system) using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer’s instructions. Cells were harvested 72 h post-transfection and the crude viral lysate was purified by iodixanol density gradient ultracentrifugation. The virus was subsequently concentrated and exchanged with buffer. Viral genome titers were determined by quantitative PCR (qPCR), using a linearized plasmid standard. All purified AAV9 preparations were formulated in 0.001% Pluronic F-68 solution (Poloxamer 188; Caisson Laboratories, Smithfield, UT, USA) for in vivo application.

Mouse models of osteoporosis induced by OVX and aging and treatment with AAV-Gpr125

Two-month-old female C57BL/6J mice were subjected to either ovariectomy (OVX) or sham surgery (removal of approximately 1 g adipose tissue adjacent to the ovaries). One month post-surgery, OVX mice received an intra-articular (I.A.) injection of 5 µL of AAV9-NC (negative control) or AAV9‐Gpr125 (titer = 1 × 10¹³ mL⁻¹) into the skull bone or femoral bone marrow cavity. The experimental timeline for OVX surgery and subsequent AAV9 treatment is shown in Fig. 9A. To evaluate the bone-targeting efficiency of the Runx2 promoter-driven AAV9‐Gpr125 construct, GFP expression in various tissues was monitored using a fluorescence molecular tomography imaging system (PerkinElmer FMT4000, USA). 23-month-old male C57BL/6J mice received an intra-articular (I.A.) injection of 5 µl of AAV9‐NC (negative control) or AAV9‐Gpr125 (titer = 1 × 10¹³ mL⁻¹) into the calvarial bone or femoral bone marrow cavity. Bone tissue was collected one month after injection. All wild-type (WT) C57BL/6J female mice used in this study were supplied by Hunan Slek Jingda Laboratory Animal Co., Ltd. All animal procedures were conducted in compliance with the guidelines and approved by the Institutional Animal Care and Use Committee of the Third Xiangya Hospital, Central South University.

H&E staining

Femurs and tibiae were isolated from mice, dissected to remove skin and soft tissues, and fixed overnight in 4% paraformaldehyde (PFA) in 1× PBS at 4 °C. The samples were then dehydrated using a graded ethanol series and decalcified in 10% (ethynediaminetetraacetic acid EDTA (; pH 7.4) for one week at room temperature with gentle agitation. For paraffin embedding, tissues were further dehydrated in ethanol, cleared in xylene, infiltrated with paraffin, and sectioned at a thickness of 6 μm using a Leica microtome (Leica Microsystems, Germany). The sections were mounted on Superfrost Plus slides (Fisher Scientific, USA). For staining, sections were deparaffinized in xylene and rehydrated in a graded ethanol series in distilled water. They were then stained with hematoxylin, differentiated in 1% acid alcohol, blued in ammonia water, counterstained with eosin, dehydrated through an ethanol series, cleared in xylene, and coverslipped with a mounting medium.

Micro-computed tomography (Micro-CT) analysis

Femurs were harvested from euthanized mice and meticulously dissected to remove all adjacent muscles and connective tissues. The cleaned bone specimens were fixed in 4% paraformaldehyde (in 0.1 M phosphate buffer, pH 7.4) for 24 h at 4 °C and subsequently rinsed in phosphate-buffered saline (PBS). The distal femoral regions were scanned using a Quantum GX2 micro-CT imaging system (PerkinElmer, USA) under the following acquisition parameters: X-ray voltage, 80 kV; current, 70 µA; field of view (FOV) = 7.68 mm, voxel size = 9.7 μm, exposure time, 4 min; rotation angle, 360°; and a total of 1440 projection images were acquired per sample. Scans were reconstructed using a filtered back-projection algorithm with the beam hardening correction set to 30%. Three-dimensional morphometric analysis was performed using the manufacturer’s software (Analyze 12.0, PerkinElmer) to quantify trabecular bone parameters in a defined region of interest (ROI) located 0.5 mm proximal to the growth plate, covering a volume of 1 mm in height. All scans and analyses were performed by blinded operators to avoid bias.

Statistical analysis

All quantitative data are presented as mean ± standard deviation (SD). Statistical analyses were performed using the GraphPad Prism software (version 8.0; GraphPad, San Diego, California, USA). Between-group comparisons were conducted using two-tailed unpaired Student’s t-tests. Statistical significance was set at p < 0.05. The following symbols denote the levels of significance: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. All experiments were independently repeated at least thrice, and the sample sizes (n) are indicated in the figure legends.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

CY.T., HD.Z., and M.L. designed this study. CY.T., YX.H., HX.W., HL.J., Y.W., Q.H., and YL.L. collected the data and performed analyses. H.L., G.F., YF.L., and M.W. provided experimental technical support. CY.T. and YS.M. performed a bioinformatic analysis. C. T., H. D. Z., and M. L. drafted the manuscript. All authors have approved the final version of the manuscript for submission.

Funding

This work was supported by the National Natural Scientific Foundation of China [grant numbers: 82300995, 82170900], Natural Science Foundation of Hunan Province (No. 2023JJ40907), Natural Science Foundation of Changsha (No. kq2208355), The “co-PI” project from The Third Xiangya hospital of Central South University (Number: 202422), National Science and Technology Major Project [grant number: 2023ZD0509205], Key Research and Development Program of Hunan [2024JK2113], Hunan Provincial Health High-Level Talent Scientific Research Project [grant number: R2023019], and Health Commission of Hunan Province Project [grant number: D202303017033].

Data availability

All data presented in the study are presented in the manuscript and supplementary files. Further inquiries of datasets of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study strictly adheres to international guidelines. Prior to the start of the experiments, we obtained approval from the Institutional Animal Care and Use Committee of the Third Xiangya Hospital of Central South University. The Animal Ethics Committee at the hospital follows the rules of Basel Declaration.

Consent to for publication

Not applicable.

Conflict of interest

The authors declare no conflict of interest.